System and Method for Extracting Deep Learning Based Causal Relation with Expansion of Training Data

摘要

The present invention is a device for extracting causal relationships based on deep learning by expanding learning data to efficiently construct a causality model in a new domain while minimizing the problem of data construction due to domain specialization when extracting causality from sentences. A data analysis and expansion unit that extends data by labeling for causality extraction by applying data analysis, applying bootstrapping, and applying active learning and bagging to alleviate error propagation; Is used as input, and the output values of forward and backward of each sequence are combined to finally derive the correct answer label for each morpheme, reflect word-of-speech unit embedding, and syllable unit embedding to reflect part-of-speech information. Causality extraction using reflected clue word dictionary qualities It intended to include; layer learning based causal extractor.